The Structure of Signed & Spoken Languages

THE STRUCTURE OF SIGNED & SPOKEN LANGUAGES Sherman Wilcox Two questions Linguistic research on American Sign Language (ASL) over the past three decades has established without question that ASL is a natural human language, distinct from English (Stokoe 1960, Stokoe et al. 1965, Klima & Bellugi 1979, Bellugi & Studdert-Kennedy 1980, Wilcox 1988b). Linguistic studies of several other indigenous signed languages have led to the same conclusion: primary sign languages, those used (mainly) by deaf people, are fully developed human languages independent of the languages spoken in the linguistic communities in the same region. These findings raise two questions. First, what is the relationship between spoken and signed languages? Are signed languages merely analogues of spoken languages, the linguistic equivalent of the bat's wing (evolved quite differently from the bird's wing)? Or are they true homologues, biologically related, as the human lung is to the swim bladder of fish? One objective of research now in progress is to frame the study of signed languages in terms that will lead to answers to the question of relationship. Specifically, it will explore a model that describes both spoken and signed languages as gestures. Second, what can the study of signed languages tell us about the human capacity for language? (See Fromkin 1990.) Linguists have not hesitated to propose theories of "human language" based on data drawn only from spoken languages. By considering all natural human languages-both signed and spoken-we can gain a better understanding of how language is represented in the human brain. (See Poizner et al. 1987 and a review by Kimura, 1988 [1990].) The model that C 1990 by Linstok Press, Inc. See note inside front cover. ISSN 0302-1475 141 Wilcox encompasses both spoken and signed languages assumes that the key lies in describing both with a single vocabulary, the vocabulary of neuromuscular activity. Speech as gestures Oddly, much more work has been done to describe speech than to describe signing in gestural terms. The psychologist Ulrich Neisser, for example, has noted that it is possible to describe speech as "articulatory gesturing, and to treat speech perception as comparable to perceiving gestures of other kinds" (1976). Michael Studdert-Kennedy, a speech researcher also with the Haskins Laboratories, has suggested that speech can be characterized as "subtly interleaved patterns of movements, coordinated across articulators" (1987). The description of speech as gestures contrasts with the traditional description of speech in terms of abstract linguistic units. In the traditional framework such units of language-segments for examplehave three properties: they are discrete, they are static, and they are context-free. Speech is consequently seen to consist of the sequential ordering of discrete states (or targets) of the vocal tract. Problems develop when we consider how these abstract units are realized as articulations, as physiological activity. First, speech as actually produced is not discrete and context-free; segments influence each other in a process known as "coarticulation." The influence has been observed to extend up to three segments in either direction; e.g. consider the word spoon. In their canonical forms, the segments /s/, /p/, and /n/ do not have lip rounding, but the segment /u/ (the "oo" sound of English) does. When spoken sequentially in the word spoon, all the segments are produced with lip rounding. Coarticulation operates in a right-to-left fashion on the /s/ and the /p/ of spoon and left-to-right on the /n/; i.e. the lips are rounded even as pronunciation of spoon begins and remain rounded until it ends, even though only the vowel segment calls for rounding (Daniloff & Hammarberg 1973). Second, the vocal tract is rarely if ever in stasis. As Carol Fowler, a speech researcher at the Haskins Laboratories observes: ". . . if segments are only achievements of vocal tract states .. . then most of the talking process involves getting to segments, and less of it is involved in actually producing them" (1987a). SLS 67 Language as gesture Problems also arise in perception. It seems that we do not always perceive static targets. Winifred Strange (1987) manipulated sequences of segments, typically consonant-vowel-consonant (CVC) sequences, to test whether perceivers rely on information for canonical "target" formant frequencies, or for transitions. Because of...